REVEAL-NOx Flight Summaries

REducing aViation Emissions’ uncertAin cLimate impacts: NOx

“The chain of events from emissions of NOx to impacts on climate are complex and uncertain. Are these aviation NOx emissions responsible for a lot or a little warming or cooling? In this project we aim to constrain the uncertainty we have on these climate impacts by constraining the performance of models in representing the chemical cascades from aviation NOx emissions” - Professor Alex Archibald, speaking to NCAS Comms

Eight flights aboard the FAAM aircraft were conducted in June 2025 (C409 - C418) following roughly the same flight plan, either clockwise or counter-clockwise.

The counter-clockwise case:

  1. Take off, heading east and climb to flight level 300
  2. Fly a circuit of off the East of England, following flight lanes L603 and L602
  3. Turn north and fly towards Aberdeen along the east coast of the UK
  4. Just south of Aberdeen, turn south and ascend to flight level 340
  5. Fly due south, roughly tracing the 2° west line of longitude until south of Birmingham
  6. Recover to Cranfield

All flights were flown as doubles, with the direction reversing in the afternoon, such that the circuit and 2° W transect were flown at both FL300 and FL340 each day. Flight tracks can be viewed below.

Overview

NOx processing details

Data + Cleaning

  • All data was used, including those flagged Reduced Quality or Suspect due to the lack of clear definition of these terms
  • Apply 10-6 - 106 pptv threshold for spike removal

The Limit of Detection (LOD)

  • The 10 Hz NOx data is often close the the limit of detection, especially when the uncertainty is considered
  • Time averaging improves this, here we average to 10 s, reducing the uncertainty \(\sqrt{\frac{\sum{u^2}}{N}}\) and LOD by \(\sqrt{\frac{\sum{LOD}}{N}}\)

The below plot demonstrates this on two 500 s periods taken from C409, data are red when the mixing ratio - uncertainty are less that the LOD (shown as a black line). The y-axis is pseudo-log10 1.

Move the slider to switch between the 10 Hz and 10 s data.

Data recover is much improved when treating the LOD in this fashion. The below plots shows 10 s data for all flights

Footnotes

  1. The psuedo-log transformer: ↩︎